True "individualization" of cancer therapy

One major point I’ve tried to make over the last few years is that the so-called “individualization” or “personalization” of treatments claimed by practitioners of “complementary and alternative medicine” (CAM) is not “individualization” at all, but rather a sham that appears superficially like individualization but in reality is not. I say that because the “individualization” promoted by CAM practitioners is not based on science and clinical trials. Another point I’ve been trying to make is that the true “individualization” of treatments will require science, and it will not be easy. In fact, it will be very, very hard.

This point was driven home over the weekend in an article by Gina Kolata in the New York Times entitled In Treatment for Leukemia, Glimpses of the Future. While the story is basically one long anecdote that shows what can be done when new genomic technologies are applied to cancer, it also shows why we are a very long way from the true “individualization” of cancer care. It also turns out that I’ve discussed the same basic story before, but here I’ll try to discuss it in a bit more detail.

As hard as it is to believe, it’s been nine months since Steve Jobs succumbed to a metastatic neuroendocrine cancer of the pancreas. Last November, the authorized biography of Steve Jobs, written by Walter Isaacson, revealed that after his cancer recurred for the second (and final) time Jobs became one of the first twenty people in the world to have all the genes of his cancer and his normal tissues sequenced, which was done by a collaboration of research teams at Stanford, Johns Hopkins, and the Broad Institute. At the time (2010-2011), it cost $100,000 to do. Scientists and oncologists looked at this information and used it to choose various targeted therapies for Jobs throughout the remainder of his life, and Jobs met with all his doctors and researchers from the three institutions working on the DNA from his cancer at the Four Seasons Hotel in Palo Alto to discuss the genetic signatures found in Jobs’ cancer and how best to target them. Jobs’ case, as we now know, was, alas, a failure. However much Jobs’ team tried to stay one step ahead of his cancer, the cancer caught up and passed whatever they could do.

Kolata’s story, in contrast, appears to be that of a success. It is the story of Dr. Lukas Wartman, a recent graduate of the Washington University hematology-oncology fellowship who is now an Instructor in the Division of Oncology:

Genetics researchers at Washington University, one of the world’s leading centers for work on the human genome, were devastated. Dr. Lukas Wartman, a young, talented and beloved colleague, had the very cancer he had devoted his career to studying. He was deteriorating fast. No known treatment could save him. And no one, to their knowledge, had ever investigated the complete genetic makeup of a cancer like his.

So one day last July, Dr. Timothy Ley, associate director of the university’s genome institute, summoned his team. Why not throw everything we have at seeing if we can find a rogue gene spurring Dr. Wartman’s cancer, adult acute lymphoblastic leukemia, he asked? “It’s now or never,” he recalled telling them. “We will only get one shot.”

We learn later in the article that Dr. Wartman had first been diagnosed with acute lymphoblastic leukemia in 2002, when he was a fourth year medical student ready to finish up and move on to residency. During trip out to California for a job interview, he began to experience overwhelming fatigue. Upon his return, he found he couldn’t run anymore and started having night sweats. At first he thought it was mono, but then he started having bone pain. He finally went to an urgent care center, where at first it was thought that he might be suffering from depression, but it was also noticed that he had a low white blood cell count.

The rest of his story can be summarized in a manner that is still too common among leukemia patients. He underwent nine months of intensive chemotherapy, which was followed by 15 months of maintenance chemotherapy. Five years later, his leukemia recurred, and he underwent a bone marrow transplantation, which is the usual treatment for recurrent ALL. Thus far, he had beaten the odds, having not realized how bad they are for recurrent ALL:

Seven months after the transplant, feeling much stronger, he went to a major cancer meeting and sat in on a session on his type of leukemia. The speaker, a renowned researcher, reported that only 4 or 5 percent of those who relapsed survived.

“My stomach turned,” Dr. Wartman said. “I will never forget the shock of hearing that number.”

In the vast majority of cases, a patient like Dr. Wartman would be dead. As is pointed out in Kolata’s article, we are not even sure of the expected survival rate of someone who has relapsed twice with ALL, other than that the odds are clearly very, very low. However, Dr. Wartman was very fortunate to have friends who had access to technologies to which few have access. Because he is a researcher at the Washington University and knew Dr. Timothy Ley, a world-renowned leukemia researcher and Associate Director for Cancer Genomics for the The Genome Institute at Washington University and because he was apparently well-liked there, Dr. Ley offered to do what was described in the introduction of Kolata’s article. It was a massive undertaking, as well:

Dr. Ley’s team tried a type of analysis that they had never done before. They fully sequenced the genes of both his cancer cells and healthy cells for comparison, and at the same time analyzed his RNA, a close chemical cousin to DNA, for clues to what his genes were doing.

The researchers on the project put other work aside for weeks, running one of the university’s 26 sequencing machines and supercomputer around the clock. And they found a culprit — a normal gene that was in overdrive, churning out huge amounts of a protein that appeared to be spurring the cancer’s growth.

Even better, there was a promising new drug that might shut down the malfunctioning gene — a drug that had been tested and approved only for advanced kidney cancer. Dr. Wartman became the first person ever to take it for leukemia.

And now, against all odds, his cancer is in remission and has been since last fall.

As is often the case when I’m reading an article in the lay press, I sometimes have to read a bit between the lines and make an educated guess as to what exactly it was that Dr. Ley’s team did. The first analysis appears to be a next generation sequencing (NGS) analysis of both the leukemic cells and normal cells. NGS techniques allow the sequencing of complete genomes in a matter of weeks. In “old-fashioned” automated sequencing using Sanger techniques, the rate-limiting step in the sequencing process was the need to separate reaction products on a polyacrylamide gel in order for the sequence to be read. Next generation sequencing (NGS) techniques overcome this limitation by arraying DNA molecules on solid surfaces by anchoring and copying single DNA molecules on glass slides or or array of beads. Going into the details of these new sequencing techniques is beyond the scope of this particular post (maybe some day), but for interested readers who know a bit about sequencing and PCR, there is a decent general description here. The long and the short of it is that NGS techniques allow the massively parallel sequencing of a genome such that 30 gigabases of DNA sequence, which is the equivalent of approximately 10 haploid human genomes, can be obtained in a week at a cost of approximately $15,000. By way of comparison, the draft human genome reference that was reported in 2001 to six-fold redundancy took five years of sequencing by several laboratories and cost billions of dollars. That is how much technology has advanced in a single decade. It’s truly astonishing.

The second analysis that was performed is almost certainly another NGS techique known as RNAseq, which is the common name for whole genome shotgun sequencing (WTSS). Again, the details of the technique are beyond the scope of this post. However, RNAseq overcomes the major limitation of cDNA microarrays, which is that it is only possible to measure the mRNAs whose sequences are known and therefore have been placed on the gene chip. Consequently, cDNA microarrays can’t discover previously unknown transcript and in general do not cover noncoding RNAs, such as microRNAs and long noncoding RNAs (lncRNAs). RNAseq does. As a result, using RNAseq it is possible to identify every sequence of every mRNA transcript, coding and noncoding, being made by the cell and how much. Of course, cDNA microarray techniques are by no means dead yet, the primary reason being that RNAseq is a lot more expensive than cDNA microarray techniques, at least ten-fold more, and cDNA microarray experiments can be completed a lot faster.

Taking the results of the sequencing of the entire genome and RNAseq data and analyzing them allows scientists to probe the genome and transcriptome of cancers in a way that was never before possible. It produces an enormous amount of data, too, terabytes from a single experiment. At cancer meetings I’ve been to, investigators frequently refer to a “firehose” of data, petabytes in magnitude. Indeed, the sheer quantity of data from these experiments challenges the bandwidth of universities doing them, and, in fact, it’s not at all uncommon for the preferred means of sending experimental data to be to load up a hard drive with the data and send it by the quaint but effective method of overnight mail to other investigators because it’s faster and more reliable that way. Not surprisingly, serious computing power and major advancements in computer algorithms have been necessary to develop the methods of analyzing data from these experiments.

What I’m trying to convey is that what WUSTL did for Dr. Wartman was not a little deal. It was a big deal that took a lot of resources and effort and likely cost well over $100,000. Apparently it was paid for through research grants, and Dr. Ley claims that no patients were neglected while all that sequencing and computing firepower were transferred to sequencing Dr. Wartman’s cancer genome and transcriptome, having done the same thing for a previous patient. That might well be true, but does anyone believe that Dr. Wartman would have had access to so much genomics goodness if he hadn’t been a researcher at The Genome Institute? Be that as it may, Dr. Wartman’s luck didn’t end at having friends willing to go to such great lengths for him. Here’s what happened when all that sequencing was done and analyzed:

The cancer’s DNA had, as expected, many mutations, but there was nothing to be done about them. There were no drugs to attack them.

But the other analysis, of the cancer’s RNA, was different. There was something there, something unexpected.

The RNA sequencing showed that a normal gene, FLT3, was wildly active in the leukemia cells. Its normal role is to make cells grow and proliferate. An overactive FLT3 gene might be making Dr. Wartman’s cancer cells multiply so quickly.

Even better, there was a drug, sunitinib or Sutent, approved for treating advanced kidney cancer, that inhibits FLT3.

In brief, for whatever reason, Dr. Wartman’s leukemia cells appeared not to have any mutations in the FLT3 gene, which would have been found in the DNA sequencing, but for some reason (probably a mutation in a regulatory region) was making lots and lots of the kinase coded for by FLT3. FLT3 has also been implicated as a molecular target in acute myeloid leukemia (AML). Indeed, this is the very reason why sequencing the genome and transcriptome both are frequently needed to understand what is driving the cancer. Of course, as I’ve discussed before, the genome of your typical cancer cell is so messed up that it’s impossible to identify a single gene that is primarily responsible for the cancer, but in this case Dr. Wartman was again incredibly lucky. His recurrent ALL was being driven primarily by FLT3, a single gene, and there exists a good drug to target that oncogene. Even better, as predicted by the biology, treating Dr. Wartman with sunitinib worked! His blood cell counts started to normalize within days, and he rapidly went into remission.

There was a hitch, however. Sunitinib is very, expensive ($330 per day), and Dr. Wartman’s insurance company wouldn’t pay for it, given that it was being used off-label. Basically, Dr. Wartman scraped together enough money to buy a week’s worth of the drug, and that’s what demonstrated such remarkable effects. Later, the doctors in his division pitched in to buy him more drug. After a few weeks, Dr. Wartman was in complete remission. Fantastic news, but it presented a dilemma: Should Dr. Wartman keep taking the drug, or should he undergo another bone marrow transplant? Ultimately, the decision was made to do another bone marrow transplant because of fear that the leukemia would soon evolve resistance. When evolution meets modern medicine, evolution nearly always wins. In another stroke of luck, Pfizer decided to supply Dr. Wartman with the drug free of charge until he underwent a bone marrow transplant. As of the running of the story, Dr. Wartman remains, as far as the best tools of modern medicine can tell, free of cancer, although he is suffering from graft versus host disease due to his transplant.

There’s no doubt that “individualized” medicine will become increasingly a part of modern medical care, with the individualization based on sequencing the genomes and transcriptomes of patients. In just a few years, the price of a complete genome sequence has fallen from hundreds of thousands of dollars to around $15,000. True, that doesn’t count all the analysis and that’s $15,000 per genome, which means at least $30,000 to sequence a normal and cancerous genome. There are, however, lots of things we do in medicine that cost $15,000. The price doesn’t have to come down much more before whole genome sequencing starts to look doable for individual patients. After all, gene tests like the OncoType DX cost on the order of $3,000 to $4,000, and we now order this test fairly routinely for patients with estrogen receptor-positive, node-negative breast cancer because in the end it saves a lot of patients from unnecessary chemotherapy.

The problem with the individualization of care based on genomics are well-illustrated in this article, which, let’s not forget, is nothing more than anecdote. The question of its generalizability remains to be determined. Using genomics to individualize treatment worked in Dr. Wartman’s case. What his odds of long-term survival are now, no one really knows, but we do know this much. There’s little doubt that, without the discovery that sunitinib would be an appropriate drug to treat his cancer, he would almost certainly be dead by now. Unfortunately, his example can be countered with that of Steve Jobs, whom sequencing his tumor ultimately didn’t help, and Christopher Hitchens, who according to this article also had his cancer sequenced.

How many will have results like Dr. Wartman’s and how many will have results like those of Steve Jobs or Christopher Hitchens? Most cancers are not driven by just one gene that can be targeted, nor are most other diseases and conditions that we might wish to use genome and transcriptome sequencing as a guide to therapy. We’re drowning in genomic data right now, and we just don’t know how to use it yet. Nor will we know until a lot more research is done. The problem is that, with a relatively few exceptions like the case of Dr. Wartman, we don’t know enough yet to translate genome and transcriptome sequences into therapies. We also don’t have drugs for anywhere near all the potential molecular targets that can be identified this way, and the targeted drugs that we do have tend to be enormously expensive. For all the promise it shows and for the now occasional success story like that of Dr. Wartman, the genomics revolution will, like most revolutions, be messy.